Variation in the adductor muscle-scar rosette of Darwinula cicatricosa Wakefield, 1994 (Ostracoda, Crustacea), and comparison with other species of Darwinula

Variation in the morphology of the adductor muscle-scar pattern of Darwinula cicatricosa Wakefield, 1994 is demonstrated, and is shown to affect specific segments within the essentially rosette-shaped scar. Potential subdivisions of individual segments can be identified in fossil specimens as ‘scar-sutures’. All fossil and Recent species of Darwinula studied are shown to have different patterns to their adductor muscle-scar rosette. These differences can be used as an aid in species identification, particularly of fossil species, in which the muscle-scars are the best indicator of soft part morphology.


INTRODUCTION
The aim of this paper is to demonstrate and explain the variation in the adductor muscle-scar (AMS) rosette of Darwin& cicatricosu Wakefield, 1994 from the Great Estuarine Group (Bathonian, Middle Jurassic) of the Inner Hebrides off NW Scotland (Figs 1 & 2). The results of this study can then be applied in the study of other species of Durwinirlu, in order to explain variation in their AMS rosettes. The AMS pattern can then be used in species identification.

THE MUSCLES AND MUSCLE-SCARS OF THE OSTRACODA
The morphology of the valves and the soft parts, in particular of the appendages and genitalia, form the basis of the classification of the Ostracoda. Excepting exceptional circumstances, e.g. phosphatized ostracods with their appendages from the Cretaceous Santana Formation of Brazil (Elate, 1972) and from the Cambrian of Sweden (Muller, 1979(Muller, , 1982(Muller, , 1990, only the valves are available to the paliieontologist. Muscle-scars represent the attachment points on the valves of the muscles of the body. The many different muscles have a variety of functions including, in the case of the adductor muscles, the closing of the valves. The morphology of the AMS is used in the taxonomy of both Recent and fossil Ostracoda. Their number, shape, distribul ion and overall pattern provide the best direct correlatio,n between the anatomy of the ostracod body and the morpi'lology of the valves. The overall pattern of the AMS is used for distinguishing various suborders, superfamilies, f,amilies iind subfamilies of ostracods. Finer details of the AMS often appear to distinguish genera, and in some cases even species (for example see van Morkhoven, 19162). Bate (1963) noted differences in the muscle-scars of the Cytheracea. He described four subtly different patterns based primarily upon variations in the devc1opmt:nl. and shape of the frontal scar. These differences did not, however, serve to identify different genera, as several genera had the same pattern.
'The body of an ostracod has many different sets of muscles. The arrangement of the muscle-scars in D. cicutricosu are shown in Fig. 3. Flexor and extensor muscles are present in all appendages, and in each podomere in order to enable independent movement. The basal podomere of each cephalic appendage is attached in the dorsal area of the valves: these attachment scars are known as the dorsal muscle-scar group. Dorsal muscle-scars were first described in Darwinida by Wakefield (1994). The basal podomeres of the mandibles are attached to the valves both dorsally, and slightly ventroanteriorly of the AMS. The adductor muscles extend from valve to valve, their attachment points forming the AMS. The adductor muscle bundle originates in the centre of the body where the striated muscle fibres connect to form a chitinous rod (Moore & Pitrat, 1961), and pass, anterior to the mid-point of the body, laterally through the duplicature. The muscle fibres diverge as they approach the chitinous inner coating of the outer lamella (van Morkhoven, 1962). Ostracod muscles are attached to the valves by chitinous bosses which are inserted into the calcareous layer of the valve (Kesling, 1951). Attachment of the muscles to the bosses is through the interlacing of muscle myofibrils and boss tonofibrils (Smith, 1968). Some muscle-scars may be faintly subdivided into several parts by slight grooves known as scar-sutures (van Morkhovcn, 1962). These scar-sutures may be formed by terminal bifurcations (Smith, 1968). These bifurcations may result, through the increase in the overall attachment area, in a greater strength of attachment.

PREVIOUS RESEARCH ON THE AMS OF DAR WINULA
The distinctive, radially arranged AMS rosette of the Darwinulacea serves to differentiate the superfamily from almost all other ostracod superfamilies. Some myodocope ostracods (see Siveter & Vannier, 1990, text-fig. 17) have similar AMS rosettes. Throughout this paper individual scars within the AMS rosette are referred to as segments, as in Fig. 3. Gramm (1976) concluded that the number of segments within the AMS rosette of Darwincda had shown no evolutionary changes, such as increases or decreases in the number of constituent segments within the rosette. Gramm's study was based on only a small number of species and is not discussed further.
Most research on the AMS of Darwinula has centred upon the type species, Durwinula stevensoni (Brady & Robertson, 1870). Wagner (unpublished doctoral thesis, 1957, University of Paris) reported of D. sreuensoni that 'the imprints of the muscle-scars were visible, they form a rosette. The number of imprints does not appear to be constant'. Wagner illustrated several different examples of AMS rosettes from D. steuensoni (Fig. 4a), three of the rosettes had 10 segments and one had 11. Triebel (1941) reported only nine segments, but he was the first to note a mandibular scar (Fig. 4b), a feature which has not subsequently been recorded. Sohn (1976Sohn ( , 1977Sohn ( , 1087 figured 19 AMS rosettes from D. steuensoni, both from live collections taken in the field and from cultured specimens. He considered that the number (9-12), shape and disposition of individual segments within the AMS rosette showed great variation, and that such variation was independent of ontogeny. He agreed with Pinto & Kotzian (1961) that the number o f segments may also vary between the right and the left valves of the same carapace, and that the number of scars was not species specific.

METHODOLOGY
It was suggested that further development of scar-sutures would result in the subdivision of a scar and an increase in the number of segments within the muscle-scar field (van Morkhoven, 1962). Keij (1973) suggested that the division of individual scars within thc basic AMS field of the bairdiidae resulted in the different patterns seen. He plotted out the AMS of several species of Harvanardiu which had four, nine and 10 constituent scars, and showed how they could be derived from the basic four-scar pattern.
I t was decided to see if the complete development of scar-sutures, resulting in segment division, could explain the variation seen in the AMS rosettes of Darwinula. The initial study was based upon a population of D. cicatricosa from the type level: Bed 7 , Kilmaluag Formation, Great Estuarine Group, Prince Charles's Point, Trotternish, Skye (bed numbering from Andrews, 1985). This population consisted of carapaces of adults, A-l and A-2 instars (Fig. 3, which had been infilled with diagenetic calcite. The specimens were mounted on glass slides using balsa cement (soluble in acetone), and the original valve material was broken away from the internal calcite mould, one valve at a time, using a steel needle. In breaking the original valves off their internal IS2 A comparison of Darwinulu species calcite moulds the adductor, frontal, mandibular and dorsal muscle-scars became visible for both valves of each carapace. !some 34 individuals were prepared in this way, from which 64 AMS rosettes were photographed using an Hitachi S-520 Scanning Electron Microscope (SEM) on llford HP5 400 ASA black-and-white film. Accurate line drawings, counts and measurements were taken of the muscle-scars directly from the photographs.
After tht: initial tabulation of data, the AMS rosettes were arranged i n order, based upon the number of segments within each rosette. The rosette with the least number of segments was chosen as a provisional standard for the species (Fig. 6a). Each segment was then labelled and its position relative to the other segments was noted. The other AMS rosettes were compared with this standard. Scarsutures were noted, and their positions compared with scars that were helieved to have subdivided (Figs. 6a & b). By careful comparison, it was possible to define idealized rosettes for each of the species on which could be marked the position of the scar-sutures, as well as guide lines demarcating the arrangement of segments relative to one another. These idealized AMS rosettes can then be compared with one another, and used in species determinations.
Line drawings of AMS rosettes discussed in the text are numbered and shaded to aid in the identification of homologous segments within only that species. The similarity in shading between species is not intended to indicate homologies between species and should not be considered as such. All of the line drawings are produced so that they appear as if they were taken from an internal view of the valve. In some cases this has meant drawings being produced as mirror images of the original drawings which were from internal moulds. In so doing the confusion, which may have arisen, in having the anterior direction of drawings from the same valve facing in different directions, has been avoided.
It was decided to further check the hypothesis of segment division and increase by studying the AMS rosettes from other species of Durwinulu. Unfortunately no other large populations of Darwinula were available from the Great Estuarine Group. However, it was possible to look at four AMS rosettes from Darwinulu pulmo Wakefield, 1994, and to compare the AMS rosettes from Scottish and English specimens of Darwinulu incurua Bate, 1967. SEM photographs and line drawings of D. steuensoni from Triebel (1941), Wagner (unpublished doctoral thesis, 1957, University of Paris) and Sohn (1976Sohn ( , 1977Sohn ( , 1987 were also used in the comparative study.  Specimens from Sohn (1976Sohn ( , 1977Sohn ( , 1987 are housed in the United States National Museum of Natural History (prefix USNM) or the Hancock Museum, Newcastle upon Tyne, England.

Material
One specimen of an A-2 instar, all other specimens (33) from A-1 instars and adults. All specimens were carapaces.

Muscle-scar variation
D. cicatricosu has a moderately variable AMS rosette with 9-15 segments (Fig. 7). The number of segments present in the AMS rosette varies not only between specimens but between valves of the same carapace. However, there does appear to be a general increase in the number of constituent segments within the rosette between A-I instars and the adults (Fig. 8). The maximum diameter of the AMS rosette increases from a mean of 100 p m in A-I instars to 120 p m in the adults. The diameter of the AMS rosette in each valve o f a carapace is almost exactly the same (Fig. 9). Within an instar stage there does not appear to be a clear correlation between increase in diameter and increase in the number of segments within the AMS rosette. A mandibular scar was developed in all of the instar stages observed. In some instances (see Fig 7a [ ) there were two mandibular scars present, the scar-suture having fully developed and thus divided the scar. There were usually two frontal scars developed, though in some specimens (Fig 7a ; both valves) the frontal scars had not developed. There was no observable ontogenetic difference between the development of the frontal and mandibular scars. This species is the only Darwinula to have been described with mandibular, frontal and dorsal scars (Wakefield. 1994).

Development of scar-sutures
Usually each AMS rosette is composed of 10 segments. Howevcr, in one instance there are only nine (Fig. 7c [LV]).
The basal segment (number S) in this specimen, however, were occasionally subdivided, though no scar-sutures were noted in specimens with undivided upper segments e.g. Fig.  7b [both valves] & d [both valves]. Segment 9 divides horizontally into two, while segment 1 divides along oblique scar-sutures into two, though in a single specimen the outer sub-segment has further subdivided ( Fig. 7f [LV]). These standard divisions do not occur in any set order. However, the division of the sub-basal segment is, as previously stated, evident in all specimens with 10 or more segments. There are some apparently random subdivisions of the segments e.g . Fig 7d. However, these subdivisions do not alter the overall positions of the scars relative to one another.

Material
Four internal moulds of valves, three left valves and one right valve, no carapaces. All specimens from Bed 3, Kildonnan Member type section, Lealt Shale Formation, Great Estuarine Group, Eigg, of Harris & Hudson (1980).

Muscle-scar variation
All of the AMS rosettes illustrated contain eleven segments (Fig. 10). There are slight differences in the shapes of individual segments, e.g. 3, 4, 6 & 10. However, there is an overall similarity between the four specimens. Unfortunately there were not enough specimens to allow a similar study to that of D. cicatricnsa.

THE AMS ROSE'ITE OF DARWINULA INCURVA BATE, 1967
Material D. incurva has been recorded from both England (Bate, 1965, 19867;Ware & Whatley, 19&0), the Inner Hebrides (Wakefield 1990(Wakefield , 1994, France (Rohr, unpublished doctoral thesis, 1976, University of Berlin) and China (Pang & Whatley, 1990). The AMS rosettes were not illustrated in many of these papers. The AMS rosettes from Bates type material, numbers lo 2274 and lo 2262 were rephotogmphed on the SEM and line drawings made from the photographs. Specimens 0s 13462 and 0s 13459 of Wakefield (1990) and from two valve fragments from respectively, Bed 21, Duntulm Formation and Beds I , 4 and 5 , Kilmaluag Formation, (3 at Estuarine Group, Laig Gorge, Eigg, of Andrews (19x5). were examined in a similar way (Fig. 11).

Muscle-scar variation
The variation in the AMS rosette of D. incurua was briefly discussed by Wakefield (1990). In most specimens the AMS rosette contains 11-13 segments, however, in one specimen ( Fig. I l a ) there are only 10 segments. Segment number 5 in this specimen has a weakly developed scar-suture which, in other specimens examined, has probably subdivided the basal segment into two sub-segments. Unfortunately this specimen was lost during transport to the British Museum (Natural History). Segment 5 may subdivide into more than two sub-segments (Fig. Ile). In one instance segments 6 & 7 have subdivided along obliquely orientated scar-sutures (Fig. llf). These have not been observed in any other specimens. The arrangement of the segments relative to one another within the AMS rosette is the same in all of the specimens studied.

Material
Line drawings of AMS rosettes from Triebel (1941) and Wagner (unpublished doctoral thesis, 1957, University of Paris) were used (Fig. 4) as well as line drawings (Fig. 12) made from the SEM photographs of Sohn (1976Sohn ( , 1977Sohn ( and 1987. Some 22 different AMS rosettes were re-drawn for this study. The number, shape and distribution of the segments within each rosette were determined. As with the other species discussed above the segments are numbered and shaded.

Muscle-scar variation
A proposed standard with nine segments is given (Fig. 13d) This rosette was determined after study of all the specimens. Variations from this 'norm' can more easily be described.
There is obviously variation in the number of segments developed between specimens (8-12), though it is not possible to say if a component of ontogeny is present within this variation as was observed in D. cicatricosa. Sohn (1987), however, stated that there was no ontogenetic variation. There is also variation between valves of the same specimen, e.g., Figs 12c & d.

Development of scar-sutures
The positions of segment division are based upon the rosette standard (Fig. 13d). Dividing lines have been drawn between the segments in order to show their position relative to one another. These relative positions are useful when attempting to identify segments which have divided as the dividing segment still remains in the same position relative to the other segments. The basal segment (number 5 ) is the site of the most regularly developed division. Segment 5 divides down its centre into two almost symmetrical sub-segments. This subdivided basal segment is present in almost all of the AMS rosettes which have 10 or more segments. The scar-suture which may develop to produce this subdivision is present in Fig. 120.
Only one other segment appears to have divided in a large number of specimens, segment 9. It subdivides into two sub-segments along a horizontal median line. The two sub-segments are usually of a rounded triangular/arrowhead shape. From the published SEM photographs and line drawings it was not possible to find a specimen with a scar-suture developed in this segment before subdivision had taken place. Figure 12h clearly shows a scar-suture developed in segment 6. This scar-suture, when fully developed, probably results in the two sub-segments seen in Fig. 1211.
Not all of the AMS rosettes studied could be compared favourably with the standard (Fig. 13d). The left valve of specimen USNM 151948 (Fig. 12b) is clearly different from the right valve (Fig. 12a) which has a subdivided segment 5  Sohn (1976Sohn ( , 1977Sohn ( & 1987. Not to scale. as described above. The AMS rosette from the left valve is completely different from that of any other AMS rosette of D. steuensoni figured. A possible interpretation of the rosette is given (Fig. 12b).

TAXONOMIC POTENTIAL OF THE AMS ROSETTE
This paper proposes that the AMS rosettes of species of Darwinula are taxonomically useful in distinguishing individual species. Though the AMS rosette does vary within a species this variation is based upon a basic pattern particular to each species, which can be determined. These basic patterns from the species discussed above are clearly different (Fig. 13). Obviously detailed studies of each species and its AMS rosette are required before this proposal can be incontrovertibly confirmed.

CONCLUSIONS
( 1 ) The AMS rosette of D. cicrrtricow varies in the number of segments which comprise the rosette between individuals, between the two valves of the same individual and between instar stages (when comparing the average number of segments developed in each instar stage and the diameter of the rosette).
(2) Variation in the number of segments between individuals can also be demonstrated for D. ( 5 ) E a c h species of Lhrwinula studied a p p e a r s t o h a v e a distinctive AEvlS rosette. Use of t h e AMS rosette c a n assist in t h e idiagnosis of species. The a u t h o r therefore disagrees with S o h n (1987) who felt t h a t t h e A M S rosette was of rlo diagnostic use. However, t h e use of any single AMS rosette t o diagnose a species is considered, in t h e light of t h e variations noted, t o be unwise.